Questions tagged [torchmetrics]

9 questions
2
votes
0 answers

Using torchmetrics with nan values

I'm, working on a a DL project and using pytorch lightning and torchmetrics. I'm using a metric that is just irrelevant for examples, which means that there exist batches for which this metric will get a NaN value. The problem is the aggregated…
1
vote
0 answers

Torchmetrics Multiclass F1Score same results as Accuracy

I am trying to log several metrics over my unbalanced classification dataset using the MetricTracker of the torchmetrics library. I find that the results for Precision Recall and F1Score always equal to the Accuracy even though it should not. A…
some_name.py
  • 777
  • 5
  • 16
0
votes
0 answers

The compute method of metric Multiclass Accuracy was called before the update method

This is my code: def _train_one_epoch(self, epoch: int) -> None: self.model.train() train_running_loss = 0.0 training_progress = 0 # len of all batches total_batches = ( len(self.dataset_config.train_data) //…
Sadegh Pouriyan
  • 157
  • 1
  • 10
0
votes
1 answer

F1Score from torchmetrics is not calculated correctly when using Average Meter pytorch

I am using AverageMeter to calculate and sum the f1 score but it doesn't calculate correctly after each epoch is finished (or after evaluation). I used torchmetric to calculate the f1core. from time import time import torch import tqdm from utils…
0
votes
0 answers

How can I calculate Jaccard Index for a batch of images?

Context: I have done a detection application. For each image, M mask detections are made. Each image has N ground truths elements. For each image M and N can be different. A mask is of shape [widht,height] I've grouped those images in batches of…
0
votes
0 answers

why is the dice function from torchvision not giving the correct result?

I am getting a different result from torchvisions dice function, than what I would expect. I am sure there is a very simple explanation for this, but I can't figure it out. import torch from torchmetrics.functional import dice preds =…
0
votes
2 answers

Torchmetrics Frechet Inception Distance Weird Behaviour

I am trying to create an FID to measure the performance of my generative models on MNIST. I provide my own feature extractor. However, in order to find the output dimension of the feature extractor you provide, torchmetrics tries to pass it a dummy…
smartstix
  • 21
  • 2
0
votes
0 answers

pytorch lightning: validation with torchmetric extremely slow

I recently tried to implement a DeepLabV3 training pipeline using pytorch lightning. I wanted to use the build-in torchmetrics.JaccardIndex as my evaluation metric. My LightningModule looks like this: import torchmetrics from pytorch_lightning…
0
votes
1 answer

TorchMetrics MultiClass accuracy for semantic segmentation

Let's use the following example for a semantic segmentation problem using TorchMetrics, where we predict tensors of shape (batch_size, classes, height, width): # shape: (1, 3, 2, 2) => (batch_size, classes, height, width) mask_multiclass_pred =…
daniel451
  • 10,626
  • 19
  • 67
  • 125